Contents lists available at ScienceDirect

Environmental Science and Policy

journal homepage: www.elsevier.com/locate/envsci

How positive is “change” in climate change? A sentiment analysisFrançois Jost⁎, Ann Dale, Shoshana SchwebelRoyal Roads University, School of Environment and Sustainability, Canada

A R T I C L E I N F O

Keywords:Sentiment analysisChangeLocal governmentClimate changeCommunityDevelopment paths

A B S T R A C T

This paper describes the results of a non-computational sentiment analysis of the word ‘change’ on the subject ofclimate change, local climate innovation and development path change. Results were obtained from interviewsconducted with local government officials in 11 communities across British Columbia: Victoria, Vancouver,Prince George, Dawson Creek, North Vancouver, Campbell River, Revelstoke, Surrey, T’Sou-ke First Nation, WestVancouver, and the Kootenay Regional Districts. The sentiment analysis was undertaken to address the lack ofagreement among climate researchers and policy-makers over the meaning of change, and to determine whetherchanges were actually occurring in their development paths. As a result, several drivers and barriers to changewere identified at the local government level. Staff quality and horizontal integration were linked to the mostpositive sentiments, whereas barriers to behavioral change and the limited pace and scale of change were as-sociated with negative sentiments.

1. Introduction

COP 21 ended with an unanticipated outcome of nation statesachieving consensus on an ambitious and bold goal of holding the in-crease in the global average temperature to well below 2 °C above pre-industrial levels (UNFCCC, 2015). Since that time there has been adisconnect between Canada’s nationally determined contributions andthe ongoing efforts to achieve that target, revealing that it was largelyaspirational, with major implementation barriers (CAT, 2018). If nationstates are serious about reaching that target, it is clear that we mustmove from business-as-usual scenarios to transformative change incurrent development paths (more available in IPCC, 2007; Moore et al.,2018).

This disconnect—between the consensus on one hand, and the im-plementation of changes needed to reach the goal on another—alsoreveals a more general confusion regarding the language of change. Apersistent challenge to transition from awareness and concern to actionremains (Moser, 2016), as well as understanding the nature of changein current development paths. This research focuses on changes takingplace in each of the assessed communities that are related to climateand the transition from their current development trajectories. Theneed to achieve a rather radical transition to address the threat to socio-ecological systems posed by climate change (Foxon et al., 2009; Mooreet al., 2018) is well known. In fact the latest IPCC Special Report (IPCC,2018) approved by governments and involving 91 authors from 40countries, clearly states that to reduce the risks associated with long-

lasting or irreversible changes, limiting global warming to 1.5 °C com-pared to 2 °C would require rapid, far-reaching and unprecedentedchanges in all aspects of society.

In spite of a global agreement on the need for change, this ‘change’will undoubtedly look different in every community; its urgency, cost,consequences, and lifespan will not necessarily align. Plans for change,from small-scale to transformative, will also inevitably face complexbarriers. Yet continuing climate science, uncertainty about natural re-source limits, and ongoing conflicts over resource developments make itclear that we need to examine our current development paths and re-frame the conversations on how to change their trajectories. We believethat we are now at a stage that the world needs “transformations of ascope, magnitude, speed, and penetration that are unprecedented inhuman history” (O’Brien, 2015: 1170).

This paper addresses what we have found to be a prerequisite forimplementing and calibrating transformative change, which is namelyreflection and dialogue on ‘change’ among those who are fighting for itsinstitutional inclusion and acceleration. The analysis of the debateabout change includes the drivers and barriers to change, past successesand innovations, and adaptation and mitigation measures put in placeby varying levels of government and/or organizations.

The research literature on the environment (e.g., Scharl et al., 2015;Sluban et al., 2014, 2015; Weichselbraun et al., 2016) and more spe-cifically on climate change change (e.g., Cody et al., 2015; Fernandezet al., 2016; Hodson et al., 2018; Holmberg and Hellsten, 2014) usingsentiment analysis or opinion mining, especially in the Canadian

https://doi.org/10.1016/j.envsci.2019.02.007Received 27 September 2018; Received in revised form 11 December 2018; Accepted 21 February 2019

⁎ Corresponding author at: Centre for Environmental Science and International Partnership (CESIP), 2005 Sooke Road, V9B 5Y2, Victoria, BC, Canada.E-mail address: francois.jost@royalroads.ca (F. Jost).

Environmental Science and Policy 96 (2019) 27–36

1462-9011/ © 2019 Elsevier Ltd. All rights reserved.

T

context, is rather small, in part because such analyses are a relativelyrecent phenomenon. Additionally, besides analyzing the methodologyemployed for the sentiment analysis per se, most research on environ-mental or climate change sentiment is focused on exploring commu-nication, from patterns to its causes and effects. There is, however, noprevious work available on the use of sentiment analysis on change forelucidating the successful and failing efforts, the drivers and barriers tochange in current development paths from the perspective of localmunicipalities. Nonetheless, a number of studies have assessed the roleand the manner in which local governments in Canada and more spe-cifically in British Columbia are introducing climate-related measuresand changes in their planning and implementation efforts (Baynhamand Stevens, 2014; Burch, 2010; Burch et al., 2014; Dale et al., 2017;Drolet, 2012; Giest and Howlett, 2013; Newell and Robinson, 2018;Moore et al., 2018; Picketts et al., 2014; Shaw et al., 2014).

In what follows, we describe the results from a six- year longitudinalresearch project, MC3: Meeting the Climate Change Challenge, whichexplores climate innovation in eleven Canadian local municipalities.Using interviews from across British Columbia, Canada, over twophases of the MC3 project (coinciding with multiple political adminis-trations), the present analysis looks to the uses of the word ‘change’ inorder to create a map of real-life change (and drivers and barriers tochange) on the ground. Though each community faces vastly differentchallenges when it comes to implementing change, we can trace fromthis map the commonalities that ought to be studied in order for Canadato move towards its target at a larger and faster scale. Detailed ob-servations on what has worked and what has not worked from theperspective of local municipalities are outlined below.

2. Methodology

This study used a mixed-methods multiple case study approach (Yin,2003), since the purpose of this research was to conduct an empiricalanalysis of a contemporary phenomenon within its real-life context, inthis case individual perspectives on the concept of change.

2.1. Case selection criteria

Case selection followed a purposive information-oriented sampling;using prospective criteria, cases were selected for both sharing somesimilar characteristics and relevant contrasting background to enrichthe discussion with a diverse range of outcomes. Two primary criteriastructured the initial selection of case studies and informed the inter-view protocol.

11 Communities were selected that exhibited leadership on adapta-

tion, mitigation, integrated adaptation/mitigation approaches,and sustainability. Examples of particularly innovative action thathad either transformed emissions pathways and/or vulnerabilityor held significant promise to do so in the future were chosen.

2 Evidence of multi-stakeholder involvement and social learning.The scale of the cases was not limited to municipal governments,thus opening up the possibility of studying compelling action at aregional scale. We selected cases where action at one scale hadbeen taken up by, or was of direct relevance to, other scales.

In addition, secondary selection criteria were incorporated to cap-ture a variety of communities demonstrating notable and diverse effortsto address climate change. The following secondary criteria were in-cluded: 1) a mix of small, medium and large communities; 2) a mix ofrural and urban; 3) communities with a long history of climate changeaction and emerging leaders; 4) a mix of resource- based and diversifiedeconomies; 5) a mix of government-led and grass-roots approaches; 6)generalizability or relevance to other communities; and 7) evidence ofsocial mobilization as a component of action.

After meeting these criteria, the final selection of communities wascompleted in close collaboration with a diverse array of civil societyand government research partners. The eleven case study communitieswere the cities of Victoria, Vancouver, Prince George, Dawson Creek,North Vancouver, Campbell River, Revelstoke, and Surrey, togetherwith T’Sou-ke First Nation, West Vancouver district municipality, andthe Kootenay Regional Districts (detailed description: mc-3.ca/case-studies). To ensure the relative comparability of the case studies androbustness of analysis and findings, a comparative five-step approachdetailed in Shaw et al. (2014) was undertaken.

In the second stage, critical case sampling was employed for theselection of three key informants per case study community to conductsemi-structured interviews, with the exception of the T’Souke firstNation where only one key interviewee was available. The total 31 keyactors -mostly local government officials (ranging from chief adminis-trative officers and city managers to city staff and councilors) and of-ficial community representatives directly involved in the design andimplementation of climate change initiatives- were selected within asubsample of participants from the first phase of MC3, 2012 to 2015(Newell and Dale, 2015; Shaw et al., 2014; www.mc-3.ca).

2.2. Data collection and analysis

The interviews were carried out using an open-ended, semi-struc-tured interview protocol that drew upon three theoretical lenses - multi-level perspectives, social practice theory, and social-ecological systems(see Moore et al., 2018). Guiding questions (15–25) were formulated todetermine the nature and degree of change that has occurred in the casestudy communities since 2012 and the way these changes includingclimate change innovations were shifting the way communities devel-oped over time. Interviewees were told in advance that their responseswould remain confidential, and their colleagues would not be aware oftheir participation in the study. These interviews which lasted from 30to 60min, were audio recorded and transcribed with permission of theparticipants. Since our present analysis focuses on the use and contextof specific language, we adapted elements from the computationalmethod of text mining in order to narrow down the body of text topassages containing the word “change.” Using the computer softwareNVivo (v11), search queries were built for the relevant codes, com-prised of words associated with the code, and included word deriva-tions and stemmed words.

Next, we applied sentiment analysis to sort the excerpts into positiveor negative groupings. Opinion mining or sentiment analysis, used in-terchangeably (Kennedy, 2012; Pang and Lee, 2008; Thelwall et al.,2011), refers to the analysis of the explicit or implicit expression—inthis case lexical—of a person to determine their attitude with respect toa topic. As in Kim and Hovy (2004), there were four evident points ofclassification: the attitude holder, the target object or aspect, the claim,and the attitude. The attitude holder is the interviewee linked to aspecific case study community. The target object is “change” related toclimate and the transition to a new development path. The claims arethe different topics, grouped into results and discussions as externallyand internally driven changes and their subdivisions, on the subject ofor concerning the target “change”. The attitude is the orientation of theclaim, being positive or negative sentiments. After removing the in-applicable excerpts and the excerpts lacking a distinct sentiment toimprove performance of the analysis (Pang and Lee, 2008), the re-maining sentiments were identified and classified according to theirpolarity.

Natural language processing (NLP) or computational linguistics are,due to the rise of social media and the web 2.0, widely employed tosystematically identify the polarity of a given text. Nevertheless, com-puter systems and NLP are not robust enough to generate accurate se-mantic representation of text (Aggarwal and Zhai, 2012). Although thelevel of inter-human disagreement on sentiment classification can be ashigh as 20% (Ogneva, 2010), this is considered to be the maximum

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

28

accuracy that sentiment classifiers based on machine learning can reach(Mozetič et al., 2016). In an effort to achieve the highest accuracy,while taking advantage of our relatively small corpus, this researchused human classification. Two separate researchers completed theanalysis with almost exact consistency. The following section providesdetailed examples from the results as well as data visualizations.

3. Results

Following Kim and Hovy’s (2004) points of classification, the results(and the following discussion) focuses on the attitudes found in relationto the different claims or issues concerning the aspect of changes inclimate or in development path. As topics around these changes variedconsiderably, with some very case specific or even considered as an-ecdotal evidence, the outcomes were assorted and summed up in 8topics and 6 subtopics. The number of positive sentiments expressedregarding change was almost double the negative ones. Fig. 1 showshow these attitudes are distributed along the different communities. Asillustrated, only two communities out of 11, namely Campbell Riverand Dawson Creek, have a higher number of negative sentiments thanpositives.

Fig. 2 shows the results obtained for the 12 most relevant topics,which are later interpreted in detail in the discussions section. As il-lustrated, only one topic presented more negative sentiments than po-sitive. Facing the increased risks of negative impacts of climate change,most interviewees related the observed and projected changes in naturewith negative sentiments.

In addition, from the analysis, and as portrayed in Fig. 2, inter-viewees expressed a similar amount of negative and positive sentimentsregarding change in relation to only two topics: mitigation measurestaken by the local government and the framing of climate change. In thefirst case, the negative sentiments reflected the discontent of re-spondents due to the limited extent and pace of local implementationefforts (insufficient change). In the second case, the negative sentimentswere related to changes in the climate change narrative, where inter-viewees argued that community members are less keen to engage indiscussions involving climate change rhetoric. Respondents mentionedthat local government tends to shift the manner of framing climatechange by addressing related issues, such as adjacent concerns or co-benefits, more relevant to community members.

Positive sentiments by official community representatives and localgovernment officials concerning changes in 10 out of the 12 major to-pics (claims) where superior or highly superior to their opposed attitude(negative sentiments). To avoid reiteration with Fig. 2, the reader isreferred to the discussions section for a detailed description and ana-lysis of these outcomes. Battery signs placed next to each of the maintopics in the discussions section are intended to visually remind thereader with the proportion of negative sentiments (in red) and positivesentiments (in green) vis-à-vis change regarding that topic.

4. Discussions

Although climate change remains a contested connotation nation-ally and internationally (Stoddart et al., 2016), this negativity has

Fig. 1. Location and distribution of sentiments on change. (Blue, positive sentiments; Yellow, negative sentiments).

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

29

decreased in recent years, these results show an overall higher numberof positive sentiments surrounding change. This higher concentration islikely due to the fact that many of the interviewees’ duties involvedtrying to introduce changes in their present organizations. Thus,changes taking place in the local government or municipalities are,more often than not, regarded by interviewees with tacit enthusiasmand positivity.

4.1. Externally-driven changes

4.1.1. Changes in natural environment

Seven case studies in this research have marine-terrestrial interfaceecosystems, considered to be highly vulnerable to climate change im-pacts (Price and Daust, 2016), from sea-level rise to more frequentextreme events such as tsunamis. Changes impacting the natural en-vironment were mentioned throughout the interviews with specialconcern for coastal storms affecting coastal wetlands. In addition, therewere frequent negative references to increases in forest fires, flooding,and the increased presence of non-native invasive species impacting thefragmented ecosystems as a consequence of climate change. Never-theless, there were positive sentiments related to changes in nature inVictoria. These were linked to increased observations of marine mam-mals, probably due to their heightened presence (Ford, 2014).

Mixed attitudes were also present regarding the perception of mi-tigation and adaptation measures taking place in the aftermath of cli-mate-related disasters. Positive sentiments were connected to increasedawareness, media coverage, and opportunities to benefit from this en-hanced awareness. In fact, extreme events often serve as a stimulus forgovernments, motivating adaptation and mitigation planning (Füssel,2007; Hamin, 2011).Whereas negative sentiments were focused on theimpermanence of the increased awareness, as well as the observedbelittlement (coming from community members and local government)of the need for urgent measures, due to the belief that risks are longer-term and carry higher levels of uncertainty. As Baynham and Stevens(2014) stated, long-term gradual changes in climatic conditions act

opposed to fast-onset events, producing a weaker reaction from gov-ernments, as they are not affected by an initial impact. This allowselected officials to focus their efforts on concrete and tangible shortterm priorities (Vogel and Henstra, 2015), which are more closelyaligned with their personal objectives during their term of office andelectoral cycle short time frames.

Another negative sentiment reported in Campbell River was theconflicting trade-off between climate action and the biodiversityagenda, especially apparent in communities with more limited re-sources. While the increased attention on climate change issues tends todivert from efforts to support endangered wildlife, biodiversity is just asurgent (Favaro et al., 2014; e-Dialogue series, 2018).

4.2. Internally-driven changes

4.2.1. Integration of the climate change agenda and intra-organizationalcollaboration

Climate policy integration is defined here as the incorporation of theaims of climate change adaptation and mitigation into other policysectors, both environmental and non-environmental (Beck et al., 2009;Dupont, 2010; Lafferty and Hovden, 2003). To enhance policy co-herence, efforts to entrench as an overarching lens or mainstream aparticular social value such as sustainability or climate change princi-ples should become institutionalized in decision-making processes (Kokand de Coninck, 2007; Vogel and Henstra, 2015).

Whether on an operational or a strategic level, changes regardingthe integration of climate change or sustainability initiatives in localgovernments were considered positive by all interviewees. Integrationimproves coordination across or between sectors, and thus rendersservices more efficient and effective (Cowell and Martin, 2003). Out-comes of the present research corroborated this for the most populatedcities, where there are greater risks in larger organizations of frag-mented or excessively entrenched silos. Representatives from the largermunicipalities, including Surrey, Victoria, Vancouver, and West Van-couver, praised the collaborative approach: “The intent is to have an

Fig. 2. Outcomes of the sentiment analysis on climate change relevant topics. Positive sentiments; Negative sentiments.

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

30

integrated model where sustainability is a core duty and responsibilityfor everybody, rather than a department” (I1, City of Victoria, tele-phone interview, 2015). Similarly, it was important for a representativein Revelstoke that the decision-making process related to climatechange not be relegated to a separate corporate function. These senti-ments align with research showing that climate change and sustainabledevelopment policies are relevant to most, if not all, departments (Yoheet al., 2007). Combining departmental initiatives with, for example,sustainability measures, has a greater effect than if implemented se-parately (Dale et al., 2017).

It is worth noting as well that it was deemed critical for smaller (i.e.population below 50,000) local governments, faced with more limitedresources, to have the help of an external expert in assisting with cli-mate change and sustainability issues, and with raising awareness.

4.2.2. Changes in leadership and staff champions

4.2.2.1. Federal and provincial governments. Four local governments(Campbell River, Vancouver, West Vancouver, and Revelstoke)emphasized the importance of increased alignment with the federaland provincial governments. Cross-scale interdependencies, namely thefederal, provincial, and local governments, tend not to be wellintegrated (Adger et al., 2005; Drolet, 2012). During the first andmost of the second phase of our research (2012–2018), provincial andlocal governments showed greater willingness to engage in climatechange efforts than the federal government (Giest and Howlett, 2013).One of the biggest drawbacks for the federal government to maintain aleadership role, as stated by Doelle (2018), was the failure to positionthe Pan Canadian Framework on Climate Change as a first step thatneeded to be gradually strengthened. Instead, the federal governmenthas taken backward steps in response to political opposition andpressure from industry sectors (Doelle, 2018). Yet multi-levelgovernance is critical to coordinate policy- making across all levels(Bauer et al., 2012; Dale et al., 2017), in order to accelerate theadoption of local innovations and to address asymmetries of scale.There is increasing pressure on local governments for more leadershipand accelerated adoption of climate innovation. Previous studies, forexample, show that pressure from higher levels of government, such astop-down directives through state or provincial mandates, provide localgovernments with motivation critical to implement strong measures foraddressing climate change (Berke et al., 1996; Homsy and Warner,2013; Tang et al., 2010).

4.2.2.2. Local government and their staff. The presence of strongleadership, both technical and political, has been identified by Burch(2010) as a key resource and crucial enabler of action on climatechange, following a study in three municipalities in British Columbia.Positive sentiments regarding changes in the local government werementioned regarding the staff number and their mindset towardsclimate action. For example, in Campbell River the highercommitment to fighting climate change since 2014 has been praisedby a respondent who considered it as very positive for boosting moraleamong the city staff, enhancing innovation, creativity, commitment,and productivity. Similarly, a Victoria representative asserted thatpolitical leadership was critical for advancing climate-relatedchanges. Positive changes in the number of staff working onsustainability, environment, and energy management were alsoreported, respectively, by interviewees from the cities of Vancouver,Prince George, and Revelstoke. A study on the incorporation of climatechange adaptation in Prince George’s community planning (Pickettset al., 2014), underlined the strength in the city staff social capital,across different sectors including planners engineers and managers.Picketts et al. (2014) stated that their previous experience andknow-how was a driver, since they were able to introduceclimate-related measures after independently proposingclimate-relevant policies.

On the negative side, the leadership changes in Dawson Creek weredeemed by a respondent as detrimental to climate innovation. The newcouncil and mayor were reported to be major oil and gas pushers andclimate change deniers, in contrast to their predecessors. In spite of this,some of the new staff in the planning department were deemed to havepassion for and interest in sustainability.

4.2.3. Changes in strategic planning

One of the most praised and positive change cited by intervieweeswas the signing of the Climate Action Charter by local governments andtheir subsequent commitment to reducing GHG emissions and takingactions around climate change. This was clearly expressed in 8 out ofthe 11 case studies (Kootenay Regional Districts, Campbell River, NorthVancouver, Prince George, Revelstoke, Surrey, Vancouver, WestVancouver), which did not mean that it was not also shared in the re-maining cases. As a result of Bill 27 amendments in 2008, local gov-ernments were required to integrate in their official community plans(OCP) GHG emission reduction targets, together with actions and po-licies for achieving those targets (Rutherford, 2009). Working as a“planning mandate”, it greatly increased the likelihood for governmentsto achieve their planning outcomes (Baynham and Stevens, 2014).

As a matter of fact, respondents mentioned that introducing thesepolicies and measures as part of their official community plans (OCP) orsustainability plans (e.g., ICSP) was a major positive change for movingforward. By incorporating climate-related mid- and long-term goals, theCharter provided a framework within which developers and the mu-nicipalities could target actions and assess development plans by fil-tering them first through climate or sustainability lenses. As stated byKootenay Regional Districts, the Charter also expanded the focus fromproviding basic infrastructure to a perspective of community well-being, including heritage, culture, and recreation.

The Charter requirements also led to increased innovation in severalcases. Respondents from Vancouver and West Vancouver mentionedthat the need for more innovative approaches influenced them to workmore collaboratively, by bringing more executive people on board andensuring full support within teams. Participation of non-state stake-holders helps ensure more effective implementation and provides anenriching, area-specific, policy-relevant knowledge and experience(Bauer et al., 2012; Agrawala and Fankhauser, 2008). In addition, thecommitment to monitoring and reporting GHG emissions had forcedlocal governments to keep track of and reduce their levels of energyconsumption, described by several respondents as a positive measure.Monitoring was also mentioned as a positive change, e.g., to bettertarget their actions (Revelstoke), or for refining target groups forhousing projects (Vancouver). However, for smaller municipalities withtighter budgets e.g., the Kootenay Regional Districts, changes promptedby the Charter were also negatively associated with increased man-agement costs required to comply with the monitoring and reportingactivities.

4.2.4. Measures implemented

A high ratio of responses within our analysis were focused on themeasures taken by local governments to target climate change andsustainability issues directly or indirectly. This should not come as asurprise, as local governments are not only directly affected by climatechange impacts but also have a substantial influence over future GHGreductions, thus locally implementing climate actions is in their bestinterests (Baynham and Stevens, 2014).

Measures such as the development of urban agriculture or com-munity gardens, where the main emphasis is on food security, weredescribed to positively influence community members by making themmore aware of their environment and more active in creating networks.Survey findings from an alternative study on B.C. communities in-cluding Prince George (Drolet, 2012) showed that respondents wereeager to support and encourage local food production. These practices

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

31

were said to contribute to building stronger and more cohesive com-munities (Drolet, 2012). Further positive changes were mentioned inKootenay Regional Districts, West Vancouver, Revelstoke, and Surrey inrelation to solid waste management initiatives, including both organicsand recycling programs, due to increased awareness among users andthe fact that they were now achieving their waste diversion goals whilereducing impacts on climate change (Ackerman, 2000). In the KootenayRegional Districts, the creation of a reserve fund to offset their corpo-rate emissions was described as a positive change.

4.2.4.1. Mitigation measures . Overall, respondents regardedmunicipal and community mitigation efforts as positive, yet turned tonegative sentiments to address gaps in the current efforts. For example,a West Vancouver interviewee spoke negatively about the absence ofprogress in monitoring or reporting on community- level GHGemissions—an action that was already implemented at the corporatelevel. They believed that implementing this at the community levelwould also encourage private enterprises and would boost overallmitigation efforts.

Regarding mitigation efforts, many spoke optimistically aboutchanges in the transportation sector. For example, three municipalitiesin the Kootenay Regional District were supporting the measurementand analysis of vehicle emissions by equipping their local bus fleets toprovide information on GHG emissions, an action that was expected tolead to more deliberative decision-making. Regarding electric vehicles,Vancouver reported that a considerable increase in the number ofcharging stations.

Efforts to reduce the use of private transportation by improving carsharing and public transport services as well as supporting activetransportation were considered positive changes in North Vancouver,Revelstoke, and Victoria. A respondent from Victoria added thatimproving infrastructure to increase the number of cyclists andpedestrians had proven to be a very successful mitigation action. Inmetro Vancouver, however, a recent survey revealed that whereas mostresidents (1/2) believed that separated bike lanes were a good thing, farfewer (1/3) wanted more in the city (Korzinski, 2018).

An additional negative sentiment involved the suburban conditionsand high-growth rate prevalent in municipalities such as Surrey, whichmade changes in emissions reduction more challenging. Similarcomments emerged for Campbell River, where the continuous effortsand initiatives made by the local government were insufficient, sincethe growth of the community was outpacing their implementationmeasures.

4.2.4.2. Changes linked to residential development . Sentimentsrelated to building regulations varied across municipalities. Vancouverinterviewees were more concerned with this subject as buildings representapproximately 65% of the city’s GHG emissions. Changes in regulations toincrease energy efficiency were highlighted, although regarded with cautionas, together with gentrification, there appeared to be a trade-off betweenthe climate agenda and the affordability agenda. The former precludeslower initial costs for construction purposes, thus conflicting with the latter.It is noteworthy to mention that housing affordability was the dominantissue of recent elections for many communities in B.C., mainly in Victoriaand Vancouver, both of which face the worst affordability conditions onrecord since the mid-1980′s, with no relief in sight (Wright and Hogue,2018).

Interviewees in Vancouver mentioned positively that thedevelopment community was open to changes, and the implementationof a future clear and longer-term government plan. Respondents fromboth Vancouver and West Vancouver agreed that the implementation ofthe building code for increasing energy efficiency in housing wassomething that they had the ability to enforce. Both were confidentabout the efforts to have new buildings built to a higher standard.Building retrofits, however, were cited as more challenging since the

local governments’ ability to encourage owners was more restricted. Yetin the Kootenay Regional District, the implementation of energyretrofits was considered by a respondent as successful.

Overall, improving the energy efficiency of new buildings in BC wasconsidered by some developers and researchers as an “easy win” or lowhanging fruit to address climate change. Describing green or zeroemission buildings as having a great potential, even without changes intechnology, and greater long term savings, covering the initialinvestment in 3–15 years (Bitz et al., 2016; Ruparathna et al., 2017)was regarded positively.

4.2.4.3. Adaptation measures . Additional changes focusingon adaptation and incorporated in provincial building regulations werealso mentioned by interviewees. Positive changes such as sea-level riseresponse were mentioned by both Vancouver and North Vancouver.Following the use of models and mapping scenarios, the minimumbuilding elevation requirement underwent a 1.1 m increase. In WestVancouver, Campbell River and Surrey, respondents encouragedchanges to take advantage of coastal engineering processes as aresponse to flood mitigation and erosive forces from wave action.Thus, instead of building dykes higher, they were seeking morebeneficial innovations such as green infrastructure or beachnourishment through offshore and barrier islands.

Faced by a different threat, Kootenay Regional District’s primaryconcern was forest fires. Consequently, proposals for new developmentsrequired land to be mitigated to FireSmart standards prior to building anew sub-division. An additional positive change was extending theinterface fire mitigation project to protect one of their drinking watertreatment facilities and infrastructure.

The reduced number of outcomes regarding changes in adaptation(see Fig. 2), including the clearly lower amount of negative sentiments,might be related to the lack of requirements for adaptations targets inthe BC Local Government Act, which left local jurisdictions withnon-binding and broad adaptation objectives (Rutherford, 2009;Picketts et al., 2014).

4.2.5. Perception of climate change in the local government

Ambiguous or conflicting preferences and the lack of convergenceamong local elected officials around particular policy solutions areconsidered a political barrier for adopting and/or implementing climatechange policy measures (Wu et al., 2018). Therefore, political will isregarded as an essential enabling condition for local climate policydevelopment (Ford and King, 2013) and critical for policy success (Postet al., 2010). Additionally Post et al. (2010: 660) divide the concept ofpolitical will into the following four components:

1 a sufficient set of decision makers;2 with a common understanding of a particular problem on the formal

agenda;3 is committed to supporting, and4 a commonly perceived, potentially effective policy solution.

Following the previous, to take a course of action decision-makersought to share a common understanding of a particular problem and itspotential effective policy solutions, in this case the perception of cli-mate change and their adaptation and mitigation measures.

Within the local government, in this analysis, climate change actionwas largely acknowledged to be insufficient, with a lack of will oftencited as the cause for inaction. Respondents from Campbell River,Kootenay Regional Districts, Dawson Creek, North Vancouver, andRevelstoke observed that changes were still limited in scope, number,aspects covered, and pace, describing the overall landscape as a tran-sition process that would take a long time to implement. The collectedresponses provide further insights about the perceived barriers and

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

32

drivers to transformative change.One positive change was the trend to recognize climate change as a

cause of natural hazards and to plan for their future consequences, asdiscussed by a respondent from Campbell River. In the KootenayRegional District, however, in spite of the recognition of climate changewithin the local government, their lower vulnerability to its impactsand thus little concern, had led to little progress on adaptation in-itiatives.

The desire to implement climate actions at a faster pace withoutcreating much chaos in the city was clearly expressed by a respondentfrom Vancouver. However, two barriers that limited their introductionwere identified. First, the lower price of fossil fuels and their deriva-tives, mad the economic benefits of renewables less evident. Second,public opinion and their acceptance of change, in this case the im-plementation of mitigation measures, was negatively related to the paceof change.

The following examples suggest that community and intra-govern-mental actions result in higher reported positivity. Despite the growthpressure, positive changes within the local government in Surrey wererelated to the willingness and ability of different departments to in-fluence other staff in favor of the relevance of climate change andsustainability. A respondent from the Kootenay Regional District posi-tively mentioned that structural issues were no longer a barrier tomoving forward on a more progressive agenda. Also, communities fromthe regional district were described as moving forward at a similar paceand sharing common goals, partially due to the regional approach theyhad taken and the sharing of knowledge and expertise between smallercommunities as a result of this scale of collaboration.

Actions by higher levels of government had mixed sentiments. Inboth the Kootenay Regional District and Victoria, there were positiveresponses to the engagement of the provincial government, the waythat they were providing guidelines and continuous feedback to thelocal government, and how renewed policies could support local in-itiatives. A respondent in Surrey considered it critical for provincialleaders to have climate change and sustainability on their agendas withthe understanding that local governments are often subject to the po-licies and mandates created by senior levels of government. In T’Sou-keFirst Nations, a respondent warned about the contradictory measuresbeing undertaken by the provincial and federal governments in thename of national economic interest and sustainable development.

Multiple respondents (Campbell River, Victoria) identified risk as amajor barrier preventing greater local government action. The fact thatthe negative consequences of inaction were expected to take place inthe future, with little probability of natural hazards greatly affectingtheir term of office, made local governments more reactive rather thanproactive, especially in the interior of B.C. Even if the personal desirewas to take adaptation measures proactively, and take the least eco-nomically disadvantageous course, the council was unlikely to supportmeasures that only pay-off in the longer term. However, the flipside ofthis was a reported positivity (in Victoria) around discussing win-winsituations, arguing that by initially focusing on positive opportunitieswith low risks, specific actions benefiting both the economic and en-vironmental agendas had a much higher chance of being implemented.

In Dawson Creek, changes in the way the local government waswilling to engage were seen as positive, although he motivation was tocapitalize on the increase in financing for green initiatives. Given thatnot only the city but also the majority of people in Northeast BC do notbelieve climate change is caused by human activities,as reported by tworespondents while quoting a recent study (Mildenberger et al., 2016),having financial benefits associated with climate-related measures iskey to persuading them to change.

4.2.6. Perception of climate change among community members

The present research also assessed community perceptions of cli-mate change-related risks and policies. According to a 2012 study in

B.C., community members do not trust the local government to effec-tively address climate change (Drolet, 2012). A literature review car-ried out by Wachinger et al. (2013), revealed that besides personalexperience, the lack of trust in authorities and experts has the mostsubstantial impact on risk perception. furthermore, a study of lack oftrust in 35 countries to blunt the effect of risk perceptions on publicwillingness to engage in behaviors or support policies to address cli-mate change (Smith and Mayer, 2018).

In this study, however, the analysis showed that community mem-bers were disengaged (due to lack of understanding and lack of will) oropposed to these policies. This lack of consensus among residents is aconsequence of the complex and intriguing behavioral intentions re-garding climate change, as people are neither complete believers orfully opposed to all personal and government efforts to adapt and re-duce GHG emissions (O’Connor et al., 1999). Results regarding disen-gagement are well supported by previous studies stating that the lack ofknowledge and low risk perceptions are a powerful predictor of beha-vioral intentions (e.g., Becker et al., 2013; Botzen et al., 2009; Dzialeket al., 2013; Lechowska, 2018; McDaniels et al., 1997; Semenza et al.,2008; Stern, 2000).

Nonetheless, most interviewees saw any effort towards involvingcommunity climate action as positive. In West Vancouver a respondentmentioned (in relation to the positive efforts of recycling) that residentswant to do the right thing, and, according to a Surrey respondent,sustainability was now the norm. In North Vancouver, people werereportedly going beyond compliance of regulations and becoming moreproactive. As Picketts et al. (2014) state, discussing climate related is-sues with the community encourages local awareness and can pre-cipitate action.

Much of the community concern for climate change was revealed tobe conditional upon the degree of direct impact. In Surrey, for example,turnout for public climate change events was strikingly low, exceptwhen these forums took place after a natural hazard, for instanceflooding near the open sea. This is consistent with the hypothesis thatgreater risk perception is associated with greater ecological, human andaesthetic impacts (Willis et al., 2005). If we take, for example, thetranstheoretical model for behavior change (Prochaska and Velicer,1997) individuals might find themselves along a temporal gradient or“stages” regarding their willingness to change or the manner in whichthey perceive climate change and are willing or not to act upon it, thuschanging their behavior.

For example, in North Vancouver and Kootenay Regional Districts,respondents described that behavioral change accelerated once climatechange affected the community in an unavoidable way. Attitudes to-wards water conservation, for example, became positive duringdroughts, with support for mandatory water restrictions. In DawsonCreek and North Vancouver, community members were engaged to thepoint of calling with complaints about the ‘other side’—those using toomuch water, or conversely oil and gas supporters who felt threatenedby new measures. A recent research on flood-related actions goes fur-ther and concludes that experiencing extreme events, rather than riskperception, represents a more significant motivation for the adoption ofprotective actions and behaviors among Canadians (Thistlethwaiteet al., 2018).

Another positive commonality was the idea of change itself as mo-mentum and a social asset. A respondent from T’sou-ke First Nationsspoke positively about making a difference with “your own hands,”instead of pleading and waiting for others to change. In Dawson Creek,one interviewee noted positively that acts of change might strengthensocial networks by bringing people closer. A positive co-benefit ofchange in Campbell River was described as acting on climate changemade people prouder of their community.

A better understanding of these interactions may be helpful for localgovernments in facilitating deliberations among decision-makinggroups on climate action policy or developing appropriate climate-re-lated communication materials for community members.

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

33

4.2.6.1. Avoiding climate change related terms while focusing on co-benefits. A significant number of interviewees pointed out that

people tend to avoid climate change discussions. Comments from theKootenay Regional Districts, Dawson Creek, Prince George and WestVancouver revealed that only negative sentiments were associated withthe climate change rhetoric, which includes talk of global warming andGHG emissions. Respondents added that climate change rhetoric wasseen as too abstract and not motivating. Therefore, local governmentswere reframing their efforts, both to councilors and staff as well as tothe wider community, in terms of the co-benefits that they bring (www.changingtheconversation.ca/capp). Similar outcomes were observed ina study with British local governments, in which 80% of the staffresponded that they have enjoyed greatest traction when climateadaptation has been rebranded as weather resiliency (Porter et al.,2015). The authors stated that focusing on weather resiliency has twoadvantages against the former, avoiding confrontation with climateskeptics and having a better buy-in as it promises to deliver immediatebenefits (as adapting to future change seems more remote).

Furthermore, respondents from the present research mentioned thatinstead of mitigation measures to reduce GHG emissions, their talkingpoints were focused on improvements to the quality of life and health,or the energy conservation and economic benefits. The latter wasespecially persuasive for implementing more mitigation and adaptationmeasures in both the public and the private sectors. As an example, themedical journal of B.C. recently published an article promoting activetransportation and framed their arguments in terms of the co-benefits,which included the reduction of GHG emissions, fewer pollutantemissions, and a more active lifestyle, in turn preventing chronic dis-ease and poor health (McVea et al., 2017).

4.2.7. Behavioral Change towards climate change

Influencing and changing the behavior of government officials aswell as community members proved challenging for all case studies,particularly in communities dependent on extractive industries such asmining, oil, or forestry. In Campbell River, for example, where manyhave depended for generations on the forestry industry, the populationwas described as very reluctant to “preserve trees”. Nonetheless, inPrince George, outcomes showed positive sentiments in the waycommunity members were more open to support sustainable initiativesonce their basic needs were met. This behavior change is supported andportrayed in Dale et al.’s (2014) community vitality pyramid, depictingthe levels or human dimensions and basic needs, having its origins inMaslow’s “hierarchy of needs” (1954).

Several interviewees discussed enforced behavioral change viapolicy changes (minor but progressive top-down shifts), however, cor-responding sentiments were mixed. Enforcing changes and lettingpeople adapt their behavior to the new system was generally describedas easier and faster to achieve meaningful change than waiting for thepopulation to change their behaviors. Respondents in Victoria, WestVancouver, and Prince George stated that proactive leadership wascrucial, in part because no amount of information can persuade somepeople to change. Others in Victoria described enforced change as apositive opportunity for creating momentum, stating that quick wins,win-win measures, and the co-benefits of acting can gradually en-courage people to be more open to change.

5. Conclusion

Much has been written on the urgent need for change to address thethreat posed by climate change, describing the role that varying levelsof government have, and their barriers and drivers for planning andimplementing climate action initiatives. Less well explored is themanner in which local policy-makers directly involved with climateaction feel about these “changes”.

Though usually applied in different domains such as natural

language processing and communication research, in this paper wepresent a new approach to sentiment analysis by identifying the con-textual polarity for a subset of sentiment expressions drawn fromtranscribed interviews conducted with decision-makers based in 11communities in British Columbia. With this approach, we are able toidentify the positive and negative sentiments of the word “change”expressed in regards to climate and the related measures carried out bylocal governments.

Subsequently, the paper discussed the causes and effects that thesesentiments implicate in the 12 most relevant topics in which these ex-pressions were grouped. The analysis revealed that positive sentimentsprevailed over negative sentiments, and many diverse examples weregiven supporting both. Whereas the majority of negative sentimentsregarded the insufficiency of the measures implemented, particularlywhen related to local growth challenges and the long path ahead toreach their GHG reduction goals, the majority of positive sentimentswere associated around institutional changes. Among the most opti-mistic were the responses to integration and collaboration inside localgovernments, and across their departments, as well as the positivechanges in leadership and staff quality.

We argue that it would be fruitful for future studies to reflect on theimplications of the present analysis. In this way, it may be possible totarget some of the areas where negative changes are taking place and todraw upon experiences in areas where positive changes were expressed.The ideas presented here should contribute to a fruitful ongoing dia-logue to further enrich the understanding of the stance of local gov-ernments vis-à-vis change and the plausible changes in developmentpaths.

For instance, support and guidance provided by the provincialgovernment was deemed crucial for local government innovation, in-cluding the changes prompted by the voluntary Climate Action Charter,providing a more level playing field. Sharing common goals has in-creased collaboration among municipalities and from external organi-zations to assist them in resolving a complex-but now common-chal-lenge. However, local governments have tended to focus on smallachievements or “easy wins” and measures that focus on co-benefits,thus limiting the scale and pace of change. This has generated frus-tration among many respondents who regarded the need to change witha sense of urgency, mentioning this was particularly influential at thepolitical level for adaptation and mitigation actions. It also reveals theimportance of communicating the urgency of acting now to the both theelectorate and elected officials, recognizing that sense of urgency isrelated to the amount of time people believe we have to act, now or inthe longer-term. How can we promote the support of local decision-makers among the diversity of municipalities and the complex politicsbetween the varying levels of government?

The restricted proactivity combined with the limited scope and paceof climate actions are both related to the limited understanding or lackof internalization of the consequences of inaction and differing gov-ernment priorities. These limitations are further affected by the mis-alignment of the provincial and federal government goals and policies.Further research is needed to understand the relevance of individualdecision-makers regarding climate action and how compliance withGHG emissions reduction goals and coordination can be ensured andtransferred to or addressed by higher levels of government.

We hope that this paper stimulates further exchanges among theinvolved staff and between local governments and that it may foster theco-production of knowledge and enhance climate change policy crea-tion inside local governments.

Declarations of interest

None.

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

34

Acknowledgements

We are grateful to the Social Sciences and Humanities ResearchCouncil (SSHRC) of Canada [Program: Insight] for the funding thatmade this research possible.

References

Ackerman, F., 2000. Waste management and climate change. Local Environ. 5 (2),223–229.

Adger, N.W., Arnell, N.W., Tompkins, E.L., 2005. Successful adaptation to climate changeacross scales. Glob. Environ. Change Part A 15 (2), 77–86.

Aggarwal, C.C., Zhai, C., 2012. An introduction to text mining. Mining Text Data.Springer, pp. 1–10.

Agrawala, S., Fankhauser, S. (Eds.), 2008. Economic Aspects of Adaptation to ClimateChange: Costs, Benefits and Policy Instruments. OECD, Paris 133p.

Bauer, A., Feichtinger, J., Steurer, R., 2012. The governance of climate change adaptationin 10 OECD countries: challenges and approaches. J. Environ. Policy Plan. 14 (3),279–304.

Baynham, M., Stevens, M., 2014. Are we planning effectively for climate change? Anevaluation of official community plans in British Columbia. J. Environ. Plann.Manage. 57, 557–587.

Beck, S., Kuhlicke, C., Görg, C., 2009. Climate Policy Integration, Coherence, andGovernance in Germany Helmholtz-Zentrum Für Umweltforschung, UFZ-Bericht 01/2009. 71p.

Becker, G., Aerts, J., Huitema, D., 2013. Influence of flood risk perception and otherfactors on risk-reducing behaviour: a survey of municipalities along the Rhine. J.Flood Risk Manag. 7, 16–30.

Berke, P., Roenigk, D., Kaiser, E., Burby, R., 1996. Enhancing plan quality: evaluating therole of state planning mandates for natural hazard mitigation. J. Environ. Plan.Manag. 39, 79–96.

Bitz, H., Butler, T., Fletcher, R., McLeod, D., Van der Star, J., 2016. The role of greenbuildings. Green Building Submission to British Columbia Climate LeadershipPlan. 8p.

Botzen, W., Aerts, J., van den Bergh, J., 2009. Willingness of homeowners to mitigateclimate risk through insurance. Ecol. Econ. 68, 2265–2277.

Burch, S., 2010. Transforming barriers into enablers of action on climate change: insightsfrom three municipal case studies in British Columbia, Canada. Glob. Environ.Change 20, 287–297.

Burch, S., Shaw, A., Dale, A., Robinson, J., 2014. Triggering transformative change: adevelopment path approach to climate change response in communities. Clim. Policy14, 467–487.

CAT (Climate Action Tracker), 2018. Climate Action Tracker Country Profile: Canada.Source URL (last Accessed 10 November 2018). https://climateactiontracker.org/countries/canada/.

Cody, E., Reagan, A., Mitchell, L., Dodds, P., Danforth, C., 2015. Climate change senti-ment on twitter: an unsolicited public opinion poll. PLoS One 10 (8), e0136092.

Cowell, R., Martin, J., 2003. The joy of joining up: modes of integrating the local gov-ernment modernisation agenda. Environ. Plann. C Gov. Policy 21 (2), 159–179.

Dale, A., Foon, R., Hebert, Y., Newell, R., 2014. Community Vitality: From Adaptation toTransformation. Fernweh Press, Tatmagouche, Nova Scotia 121p.

Dale, A., Burch, S., Robinson, J., 2017. Multi-level governance of sustainability transitionsin Canada: policy alignment, innovation, and evaluation. In: Hughes, S., Chu, E.,Mason, S. (Eds.), Climate Change in Cities: Innovations in Multi-Level Governance.Springer.

Doelle, M., 2018. Decades of Climate Policy Failure in Canada: Can We Break the ViciousCycle? SSRN. 7p.. Source URL (last Accessed 20 November 2018). https://doi.org/10.2139/ssrn.3228743.

Drolet, J., 2012. Climate change, food security, and sustainable development: a study oncommunity-based responses and adaptations in British Columbia, Canada.Community Dev. 43 (5), 630–644.

Dupont, C., 2010. Political Commitment to Climate Policy Integration at EU Level: TheCase of Biodiversity Policy. Edinburgh Europa Paper Series 2010/05. 19p.

Dzialek, J., Biernacki, W., Bokwa, A., 2013. Impact of social capital on local communities’response to floods in Southern Poland. Commun. Environ. Disaster Risk Manag. 14,185–205.

e-Dialogue series, 2018. Biodiversity Conversations. Women for Nature, Nature Canada,and Changing the Conversations. BC, Canada, September 2017 to April 2018. SourceURL (last Accessed 25 September 2018). http://changingtheconversation.ca/past-conversations.

Favaro, B., Claar, D., Fox, C., Freshwater, C., Holden, J., Roberts, A., UVic ResearchDerby, 2014. Trends in extinction risk for imperiled species in Canada. PLoS One 9(11), e113118.

Fernandez, M., Piccolo, L., Maynard, D., Wippoo, M., Meili, C., Alani, H., 2016. Talkingclimate change via social media: communication, engagement and behaviour.Proceedings of the 8th ACM Conference on Web Science 85–94.

Ford, J., 2014. Marine mammals of British Columbia. Royal BC Museum Handbook.Victoria, British Columbia. 464p.

Ford, J., King, D., 2013. A framework for examining adaptation readiness. Mitig. Adapt.Strateg. Glob. Change 20, 505–526.

Foxon, T.J., Reed, M.S., Stringer, L.C., 2009. Governing long-term social-ecologicalchange: what can the adaptive management and transition management approacheslearn from each other? Environ. Pol. Gov. 19, 3–20.

Füssel, M., 2007. Adaptation planning for climate change: concepts, assessment ap-proaches, and key lessons. Sustain. Sci. 2, 265–275.

Giest, S., Howlett, M., 2013. Multi-level governance and place-based policy-making forclimate change adaptation: the European experience and lessons for BritishColumbia. BC Stud. 176 (133-145), 147–154 199.

Hamin, E., 2011. Integrating adaptation and mitigation in local climate change planning.In: Ingram, G., Hong, Y.-H. (Eds.), Climate Change and Land Policies. LincolnInstitute of Land Policy, Cambridge, MA, pp. 122–143.

Hodson, J., Dale, A., Petersen, B., 2018. The Instagram #climatechange hashtag com-munity: does it impact social capital and community agency? Int. J. Interdiscip.Environ. Stud. 12 (3), 17–35.

Holmberg, K., Hellsten, I., 2014. Analyzing the climate change debate on Twitter: contentand differences between genders. Proceedings of the 2014 ACM Conference on WebScience 287–288.

Homsy, G., Warner, M., 2013. Climate change and the co-production of knowledge andpolicy in Rural USA communities. Sociol. Ruralis 53 (3), 291–310.

IPCC, 2018. Summary for policymakers. In: Masson-Delmotte, V., Zhai, P., Pörtner, H.O.,Roberts, D., Skea, J., Shukla, P.R., Pirani, A., Moufouma-Okia, W., Péan, C., Pidcock,R., Connors, S., Matthews, J.B.R., Chen, Y., Zhou, X., Gomis, M.I., Lonnoy, E.,Maycock, T., Tignor, M., Waterfield, T. (Eds.), Global Warming of 1.5°C. An IPCCSpecial Report on the Impacts of Global Warming of 1.5°C Above pre-IndustrialLevels and Related Global Greenhouse Gas Emission Pathways, in the Context ofStrengthening the Global Response to the Threat of Climate Change, SustainableDevelopment, and Efforts to Eradicate Poverty. World Meteorological Organization,Geneva, Switzerland 32p.

IPCC (Intergovernmental Panel on Climate Change), 2007. Climate change 2007: miti-gation of climate change. In: Metz, B., Davidson, O.R., Bosch, P.R., Dave, R., Meyer,L.A. (Eds.), Working Group III Contribution to the Fourth Assessment Report of theIntergovernmental Panel on Climate Change. Cambridge University Press,Cambridge, UK.

Kennedy, H., 2012. Perspectives on sentiment analysis. J. Broadcast. Electron. Media 56(4), 435–450.

Kim, S., Hovy, E., 2004. Determining the sentiment of opinions. Proceedings of theInternational Conference on Computational Linguistics (COLING) 1367–1373.

Kok, M., de Coninck, H., 2007. Widening the scope of policies to address climate change:directions for mainstreaming. Environ. Sci. Policy 10, 587–599.

Korzinski, D., 2018. Bike Lane Divide: Canadians more Likely to Blame Cyclists thanDrivers for Conflict on the Roads. 15p. Source URL (last Accessed 28 June 2018).Angus Reid Institute. http://angusreid.org/bike-lanes/.

Lafferty, W., Hovden, E., 2003. Environmental policy integration: towards an analyticalframework. Env. Polit. 12 (3), 1–22.

Lechowska, E., 2018. What determines flood risk perception? A review of factors of floodrisk perception and relations between its basic elements. Nat. Hazards 26.

Maslow, A., 1954. Motivation and Personality. Harper, New York, NY 369p.McDaniels, T., Axelrod, L., Cavanagh, N., Slovic, P., 1997. Perception of ecological risk to

water environments. Risk Anal. 17, 341–352.McVea, D., Peterson, E., Lu, J., 2017. How people travel daily from one place to another

has significant consequences on their health and the health of their community. BCMed. J. 59 (9), 449 486 COHP.

Mildenberger, M., Howe, P., Lachapelle, E., Stokes, L., Marlon, J., Gravelle, T., 2016. Thedistribution of climate change public opinion in Canada. PLoS One 11 (8), e0159774.

Moore, A., King, L., Dale, A., Newell, R., 2018. Toward an integrative framework for localdevelopment path analysis. Ecol. Soc. 23 (2), 13. https://doi.org/10.5751/ES-10029-230213.

Moser, S., 2016. Reflections on climate change communication research and practice inthe second decade of the 21st century: what more is there to say? Wiley Interdiscip.Rev. Clim. Change 7, 345–369.

Mozetič, I., Grčar, M., Smailović, J., 2016. Multilingual twitter sentiment classification:the role of human annotators. PLoS One 11 (5), e0155036.

Newell, R., Dale, A., 2015. Meeting the climate change challenge (MC3): the role of theinternet in climate change research dissemination and knowledge mobilization.Environ. Commun. 9 (2), 208–227.

Newell, R., Robinson, J., 2018. Using decomposition methodology to gain a better un-derstanding of progress in and challenges facing regional and local climate action. J.Clean. Prod. 197, 1423–1434.

O’Brien, K., 2015. Political agency: the key to tackling climate change. Science 350(6265), 1170–1171.

O’Connor, R., Bord, R., Fisher, A., 1999. Risk perceptions, general environmental beliefs,and willingness to address climate change. Risk Anal. 19, 461–471.

Ogneva, M., 2010. How Companies Can Use Sentiment Analysis to Improve TheirBusiness. Source URL (last Accessed 25 September 2018). https://mashable.com/2010/04/19/sentiment-analysis.

Pang, B., Lee, L., 2008. Opinion mining and sentiment analysis. Found. Trends Inf.Retrieval 2 (1–2), 1–135.

Picketts, I., Déry, S., Curry, J., 2014. Incorporating climate change adaptation into localplans. J. Environ. Plan. Manag. 57 (7), 984–1002.

Porter, J., Demeritt, D., Dessai, S., 2015. The right stuff? Informing adaptation to climatechange in British local government. Glob. Environ. Change 35, 411–422.

Post, L., Raile, A., Raile, D., 2010. Defining political will. Polit. Policy 38 (4), 653–676.Price, K., Daust, D., 2016. Climate Change Vulnerability of BC’s Fish and Wildlife: First

Approximation. BC Ministry of Forests, Lands, and Natural Resource Operations, B.C41p.

Prochaska, J., Velicer, W., 1997. The transtheoretical model of health behavior change.Am. J. Health Promot. 12, 38–48.

Ruparathna, R., Hewage, K., Sadiq, R., 2017. Rethinking investment planning and opti-mizing net zero emission Buildings. Clean Technol. Environ. Policy 19, 1711–1724.

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

35

Rutherford, S., 2009. Bill 27: Opportunities and Strategies for Green Action by BC LocalGovernments. West Coast Environmental Law Research Foundation, Vancouver,Canada 20p.

Scharl, A., Herring, D., Rafelsberger, W., Hubmann-Haidvogel, A., Kamolov, R., Fischl, D.,Fols, M., Weichselbraun, A., 2015. Semantic systems and visual tools to supportenvironmental communication. IEEE Syst. J. 11 (2), 762–771.

Semenza, J., Hall, D., Wilson, D., Bontempo, B., Sailor, D., George, L., 2008. Publicperception of climate change: voluntary mitigation and barriers to behavior change.Am. J. Prev. Med. 35, 479–487.

Shaw, A., Burch, S., Kristensen, F., Robinson, J., Dale, A., 2014. Accelerating the sus-tainability transition: exploring synergies between adaptation and mitigation inBritish Columbian communities. Glob. Environ. Change 25, 41–51.

Sluban, B., Smailović, J., Juršič, M., Mozetič, I., Battiston, S., 2014. Community sentimenton environmental topics in social networks. Proceeding of the Tenth InternationalConference on Signal-Image Technology & Internet-Based Systems 376–382.

Sluban, B., Smailović, J., Battiston, S., Mozetič, I., 2015. Sentiment leaning of influentialcommunities in social networks. Comput. Soc. Netw. 2 (9), 21.

Smith, E., Mayer, A., 2018. A social trap for the climate? Collective action, trust andclimate change risk perception in 35 countries. Glob. Environ. Change 49, 140–153.

Stern, P., 2000. Towards a coherent theory of environmentally significant behavior. J.Soc. Issues 56, 407–424.

Stoddart, M., Haluza-DeLay, R., Tindall, D., 2016. Canadian news media coverage ofclimate change: historical trajectories, dominant frames, and international compar-isons. Soc. Nat. Resour. 29 (2), 218–232.

Tang, Z., Brody, S., Quinn, C., Chang, L., Wei, T., 2010. Moving from agenda to action:evaluating local climate change action plans. J. Environ. Plan. Manag. 53 (1), 41–62.

Thelwall, M., Buckley, K., Paltoglou, G., 2011. Sentiment in Twitter events. J. Am. Soc.Inf. Sci. Technol. 62 (2), 406–418.

Thistlethwaite, J., Henstra, D., Brown, C., Scott, D., 2018. How flood experience and risk

perception influences protective actions and behaviours among Canadian home-owners. Environ. Manag. 61, 197–208.

UNFCCC, 2015. Adoption of the Paris Agreement. Report No. FCCC/CP/2015/L.9/Rev.1.Source URL (last Accessed 10 November 2018). http://unfccc.int/resource/docs/2015/cop21/eng/l09r01.pdf.

Vogel, B., Henstra, D., 2015. Studying local climate adaptation: a heuristic researchframework for comparative policy analysis. Glob. Environ. Change 31, 110–120.

Wachinger, G., Renn, O., Begg, C., Kuhlicke, C., 2013. The risk perception Paradox -Implications for governance and communication of natural hazards. Risk Anal. 33,1049–1065.

Weichselbraun, A., Scharl, A., Gindl, S., 2016. Extracting opinion targets from environ-mental web coverage and social media streams. Proceedings of the 49th HawaiiInternational Conference on System Sciences. pp. 1040–1048.

Willis, H., DeKay, M., Fischhoff, B., Morgan, M., 2005. Aggregate, disaggregate, andhybrid analyses of ecological risk perceptions. Risk Anal. 25, 405–428.

Wright, C., Hogue, R., 2018. Housing Trends and Affordability. September 2018. 10p.Source URL (last Accessed 10 November 2018). Royal Bank of Canada (RBC)Economics. http://www.rbc.com/economics/economic-reports/pdf/canadian-housing/house-sep2018.pdf.

Wu, X., Ramesh, M., Howlett, M., Fritzen, S., 2018. The Public Policy Primer: Managingthe Policy Process. Routledge, London/ New York.

Yin, R., 2003. Case Study Research: Design and Methods, 3rd ed. Sage, Thousand Oaks,CA 181p.

Yohe, G., Lasco, R., Ahmad, Q., Arnell, N., Cohen, S., Hope, C., Janetos, A., Perez, R.,2007. Perspectives on climate change and sustainability. In: Parry, M., Canziani, O.,Palutikof, J., van der Linden, P., Hanson, C. (Eds.), Climate Change 2007: Impacts,Adaptation and Vulnerability. Contribution of Working Group II to the FourthAssessment Report of the Intergovernmental Panel on Climate Change. CambridgeUniversity Press, Cambridge, UK, pp. 811–841.

F. Jost, et al. Environmental Science and Policy 96 (2019) 27–36

36